Discriminative Autoencoders for Small Targets Detection

Razakarivony, Sébastien; Jurie, Frédéric

doi:10.1109/icpr.2014.607

Cited by 24 publications

(9 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In similar work by [80], discriminative autoencoders aim at learning low-dimensional discriminative representations for positive (X + ) and negative (X − ) classes of data. The discriminative autoencoders build a latent space representation under the constraint that the positive data should be better reconstructed than the negative data.…”

Section: Controlling Reconstruction For Anomaly Detectionmentioning

confidence: 99%

An Overview of Deep Learning Based Methods for Unsupervised and Semi-Supervised Anomaly Detection in Videos

Kiran

Thomas²,

Parakkal³

2018

J. Imaging

382

153

View full text Add to dashboard Cite

Abstract:Videos represent the primary source of information for surveillance applications. Video material is often available in large quantities but in most cases it contains little or no annotation for supervised learning. This article reviews the state-of-the-art deep learning based methods for video anomaly detection and categorizes them based on the type of model and criteria of detection. We also perform simple studies to understand the different approaches and provide the criteria of evaluation for spatio-temporal anomaly detection.

show abstract

Section: Controlling Reconstruction For Anomaly Detectionmentioning

confidence: 99%

An Overview of Deep Learning Based Methods for Unsupervised and Semi-Supervised Anomaly Detection in Videos

Kiran

Thomas²,

Parakkal³

2018

J. Imaging

382

153

View full text Add to dashboard Cite

show abstract

“…Prior to the development of deep learning, a sliding window detector [ 8 ] was widely used in object detection. Sliding window methods utilize both specific hand-crafted feature representations such as HOG and classifiers such as a support vector machine (SVM) to independently binary classify all sub-windows of an image as belonging to an object or background [ 15 , 16 ]. Even though their methods have made some improvements, hand-crafted features are insufficient to separate vehicles from complex background.…”

Section: Related Workmentioning

confidence: 99%

An Improved FBPN-Based Detection Network for Vehicles in Aerial Images

Wang

2020

Sensors

View full text Add to dashboard Cite

With the development of artificial intelligence and big data analytics, an increasing number of researchers have tried to use deep-learning technology to train neural networks and achieved great success in the field of vehicle detection. However, as a special domain of object detection, vehicle detection in aerial images still has made limited progress because of low resolution, complex backgrounds and rotating objects. In this paper, an improved feature-balanced pyramid network (FBPN) has been proposed to enhance the network’s ability to detect small objects. By combining FBPN with modified faster region convolutional neural network (faster-RCNN), a vehicle detection framework for aerial images is proposed. The focal loss function is adopted in the proposed framework to reduce the imbalance between easy and hard samples. The experimental results based on the VEDIA, USCAS-AOD, and DOTA datasets show that the proposed framework outperforms other state-of-the-art vehicle detection algorithms for aerial images.

show abstract

“…There are few prior works which propose modifications to standard AE design to improve inter-class discrimination amongst features. Authors in [25] introduced a discriminative AE by combining the HOG (Histogram of Gradients) based feature selection with manifold learning. Their discriminative AE structure works on top of linear SVM classifier built using HOG features; thus requiring as many discriminative AE as the number of SVM classifiers in stage 1.…”

Section: Related Workmentioning

confidence: 99%

“…Their discriminative AE structure works on top of linear SVM classifier built using HOG features; thus requiring as many discriminative AE as the number of SVM classifiers in stage 1. In addition to being a complex design, use of HOG in [25] requires heuristic parameter selection (such as window size) and deep insight into the image. Our model, being fully automated and based on representation learning, requires no knowledge of image structure.…”

Section: Related Workmentioning

confidence: 99%

Discriminative Autoencoder for Feature Extraction: Application to Character Recognition

Gogna

Majumdar

2018

Neural Process Lett

View full text Add to dashboard Cite

Conventionally, autoencoders are unsupervised representation learning tools. In this work, we propose a novel discriminative autoencoder. Use of supervised discriminative learning ensures that the learned representation is robust to variations commonly encountered in image datasets. Using the basic discriminating autoencoder as a unit, we build a stacked architecture aimed at extracting relevant representation from the training data. The efficiency of our feature extraction algorithm ensures a high classification accuracy with even simple classification schemes like KNN (K-nearest neighbor). We demonstrate the superiority of our model for representation learning by conducting experiments on standard datasets for character/image recognition and subsequent comparison with existing supervised deep architectures like class sparse stacked autoencoder and discriminative deep belief network.The basic building blocks of these deep architectures are either the stochastic RBM's (Restricted Boltzmann Machines) [16] or the deterministic Autoencoders [17]. Given a training dataset, RBM tries to learn the network weights such that the similarity between the projection (of the training data) and the learned representation is maximized. Autoencoders (AE) on the other hand consists of two networks. The first one maps the input (training data) to the representation / feature space; the second network maps the representation space to the output (training data). Thus, an AE approximates an Identity operator; which may sound trivial, but by constraining the nodes or connections of the networks one can learn interesting representations of the data.RBM and AE are shallow architectures. Proponents of deep learning believe that better (compact / abstract) representation can be learnt by going deeper. However, learning the network weights for several layers is a difficult task. Usually, there is not enough data, the network overfits and loses its generalization ability thereby yielding subpar results at operational stage. In [17], authors presented a greedy mechanism to train the multilayer (stacked) architectures wherein each of the layer is individually trained to yield best possible representation which in turn acts as input to subsequent layer. Greedy approach learns only one network at a time, it has fewer parameters to learn, so even with limited training data, it yields better results during operation.

show abstract

Discriminative Autoencoders for Small Targets Detection

Cited by 24 publications

References 36 publications

An Overview of Deep Learning Based Methods for Unsupervised and Semi-Supervised Anomaly Detection in Videos

An Overview of Deep Learning Based Methods for Unsupervised and Semi-Supervised Anomaly Detection in Videos

An Improved FBPN-Based Detection Network for Vehicles in Aerial Images

Discriminative Autoencoder for Feature Extraction: Application to Character Recognition

Contact Info

Product

Resources

About